Skip to content
Bulka Labs

Boutique engineering consultancy

Technical decisions you won’t have to undo.

We help companies plan, build and fix the systems their business depends on. Every engagement is led by a principal engineer with more than 25 years of experience building large-scale systems. You can start with a one-hour call or a short review, and keep us on for as long as it’s useful.

Book a 30-minute callHow we work together

We work remotely and take on a few clients at a time.

Readout

Illustrative

Scale and reliability review

Findings

  • HighCheckout calls three services in sequence. One slow dependency stalls every order.
  • HighNo load test has gone past 2x current peak.
  • MediumMost alerts have no owner and no runbook.

Roadmap

  1. 30 days

    Timeouts and async calls on checkout

  2. 60 days

    Load test at 3x peak

  3. 90 days

    Failover game day

years building systems that run in production
25+
of API calls a day through a global auth runtime
Billions
concurrent viewers on a streaming platform at peak
10M+
restaurant locations on a re-architected order pipeline
80,000+

Practice areas

Where we help

Any of these can start as a fixed-fee review, a build project or an embedded engagement.

Architecture and system design

Design a new platform your team can build on, or find out whether the current one will hold up as you grow.

  • Design a new platform or service
  • Review an existing architecture
  • Plan the move off a monolith or an aging platform
  • Make a build-vs-buy or vendor decision

Ways to start

Architecture design2–3 weeks

You’re about to commit a team to a new system and want the design right before the first sprint.

For
Teams about to build a new platform or major service
You get
Design document (components, data flows, failure modes, build sequence) and a review session with your team
Timeline
2–3 weeks, fixed fee
Legacy modernization roadmap2–4 weeks

The platform slows every release, and a rewrite feels too risky to start.

For
Companies stuck on a monolith or an aging platform
You get
Migration or re-platform plan with sequencing, risk and milestones
Timeline
2–4 weeks, fixed fee
Build vs. buy and vendor selection1 week

A big platform decision is on the table and every vendor says yes.

For
Teams facing a platform or vendor decision
You get
Neutral evaluation and a written recommendation
Timeline
1 week, fixed fee

AI, RAG and agents

Find the AI projects worth doing, then build assistants and agents that work reliably with real users.

  • Rank use cases and decide what to build or buy
  • Assistants that answer from your documents, tickets and knowledge base (RAG)
  • Agents that carry out steps in your workflows
  • Evaluation, guardrails and keeping model costs under control

Ways to start

AI and agent strategy, with a build plan1–2 weeks

You’re under pressure to "add AI" and need to know where it creates value, what it takes to run in production and what to buy instead of build.

For
Product and engineering leaders deciding where AI agents belong
You get
Ranked use cases, a reference architecture, risks and a build-or-buy recommendation
Timeline
1–2 weeks, fixed fee
AI assistant or agent pilot4–8 weeks

You know where an assistant or agent would help and need one that works with real users.

For
Teams with a clear AI use case and nobody who has shipped one
You get
A working retrieval assistant or agent on your data, with an evaluation suite, guardrails, cost tracking and handover
Timeline
4–8 weeks, fixed scope and fee

Scale and reliability

Stop recurring incidents and get ready for the next jump in traffic.

  • Bottleneck and failure-mode analysis
  • Load testing for launches and peak events
  • Observability and alerts your team acts on
  • Incident response and on-call practices

Ways to start

Scale and reliability review2 weeks

Incidents keep coming back, and the next stage of growth will make them worse.

For
Teams with recurring outages or growth pain
You get
Bottleneck and failure-mode analysis with prioritized fixes
Timeline
2 weeks, fixed fee
Cloud cost reduction review1–2 weeks

Cloud spend is growing faster than revenue and nobody owns the fix.

For
Engineering and finance leaders with a growing cloud bill
You get
A list of waste, a right-sizing plan and architecture changes ranked by savings
Timeline
1–2 weeks, fixed fee

Identity, security and enterprise readiness

Pass enterprise security reviews instead of losing deals over SSO, multi-tenancy or compliance gaps.

  • SSO, OAuth 2.0 / OIDC and IAM design
  • Multi-tenancy and data isolation
  • Compliance preparation, including FedRAMP
  • SLAs and the operations needed to meet them

Ways to start

Enterprise readiness review1–2 weeks

Large customers stall in security review over multi-tenancy, SSO, compliance or SLAs.

For
B2B startups moving upmarket
You get
Gap list against enterprise requirements with a prioritized plan
Timeline
1–2 weeks, fixed fee

Expert calls

Talk it through with a principal engineer

We offer paid one-hour calls for investors, operators and engineering leaders who need a clear answer before a decision, a deal or a board meeting.

  • Investors and analysts

    • Technology risk in a deal
    • Whether a vendor’s claims hold up
    • How the platforms in a market are built and where they tend to fail

  • Founders and operators

    • An architecture or vendor choice
    • A sanity check on an AI plan
    • What enterprise buyers will ask for

  • Engineering leaders

    • A second opinion on a design
    • An incident that keeps coming back
    • What to look for in a senior hire

Topics

  • Identity and access management at cloud scale (IAM, OAuth 2.0 / OIDC, SSO, federation)
  • Streaming video platforms and readiness for peak events
  • Subscription billing, payments and fraud prevention
  • Online ordering and restaurant technology platforms
  • Distributed systems and event-driven architecture
  • Cloud infrastructure, Kubernetes and multi-region AWS
  • AI agents and RAG in production, including build vs. buy, evaluation and cost
  • Reliability, incident management and observability
  • Enterprise readiness and FedRAMP
  • Technical due diligence and engineering team assessment

Format

  • 60 minutes by video
  • Paid in advance
  • A written summary of the recommendations afterward

The free intro call is for scoping a project. An expert call is a working session on your own question.

Calls never cover confidential information about current or former employers or their customers.

Ask about an expert call

How we work together

Start small and grow from there

Every engagement starts with a free 30-minute intro call to scope the work. After that, you choose the type of engagement that fits.

  • Fixed-fee review

    1–4 weeks · Fixed fee

    We assess your system or design a new one, then deliver a written report, a 30/60/90-day roadmap and a readout call. The plan is yours to carry out with or without us.

  • Build project

    4–8 weeks · Fixed scope and fee

    We build a defined piece of work, such as an AI assistant pilot, test it and hand it over to your team.

  • Embedded principal engineer

    3–6 months · Monthly fee

    We join your team remotely to ship an agreed outcome and mentor your engineers along the way.

  • Fractional retainer

    Ongoing · Monthly fee

    We help with architecture decisions, design reviews and hiring on an ongoing basis, so you don’t need a full-time hire.

For investors and acquirers

Technical due diligence

1–2 weeks · Fixed fee

You need to know whether the target’s architecture, team and tech debt can carry the plan you’re underwriting.

You get: A risk-rated findings report, a remediation cost estimate and a readout call

Need a single call instead? Book an expert call.

Prior experience

Our founder’s track record

Our founder held these staff and systems engineering roles as an employee. The companies listed were employers, not clients.

Olo

Staff Software Engineer, 2021–2023

  • Led the re-architecture of order processing for one of the largest restaurant SaaS platforms (80,000+ locations). Raised order success rate by a full point and gave the team end-to-end visibility into every order.
  • Moved monolithic business logic into BPMN state machines so order and payment flows could be audited and recovered.
  • Built Kafka and Airflow pipelines for real-time analytics across the platform.

HBO (WarnerMedia)

Staff Software Engineer, 2013–2021

  • Core architect of the streaming platform behind HBO GO, HBO NOW and the HBO Max launch, with 10M+ concurrent viewers at peak.
  • Designed the async purchase and payment system (subscription billing, invoicing, fraud prevention) handling millions of transactions a day.
  • Built the identity and access management services for millions of accounts with federated authentication.

Amazon Web Services

Systems Engineer, IAM, 2010–2013

  • Owned core components of AWS IAM, the authentication and authorization backbone for every AWS service.
  • Engineered high-throughput messaging and tiered propagation for the global auth runtime, handling billions of API calls a day.
  • Key contributor to planning and launching the Sydney region, one of the first stand-alone AWS regions outside the US.

About

Vasiliy Boulytchev

Founder and principal consultant, Bulka Labs

Vasiliy Boulytchev is a principal engineer with more than 25 years of experience building and leading the systems behind products used by tens of millions of people, including AWS IAM, HBO’s streaming platform and Olo’s ordering pipeline. Vasiliy leads every engagement personally and is responsible for its architecture and technical direction. Vasiliy has also built production AI systems, including RAG pipelines, LLM-based diagnostics and agents that work with real operational data.

Selected capabilities

  • Architected an AI platform on self-hosted LLMs that reasons over schema, query patterns and production telemetry to diagnose issues and recommend fixes, replacing legacy BI dashboards and support tooling.
  • Built RAG pipelines over database schemas and documentation that generate troubleshooting guides and schema recommendations.
  • Led a FedRAMP migration of a cloud service, including the architecture changes, controls and evidence needed for a compliant environment.
  • Established incident management for an SLA-bound cloud service and led critical recoveries that improved uptime.
  • Owned the technical roadmap for a cloud data platform (storage, replication, AI-assisted operations) and mentored senior engineers on LLM integration and agents.

B.S. Computer Science, Colorado Technical University

Toolbox

AI
LLM integration, RAG pipelines, agent orchestration, self-hosted LLM inference
Identity and security
OAuth 2.0 / OIDC, IAM, federated authentication, SAML, FedRAMP
Data and messaging
Kafka, NATS, RabbitMQ, SQS/Kinesis, Airflow, event-driven architecture, gRPC/Protobuf
Infrastructure
Kubernetes, AWS, GCP, Terraform, CI/CD, OpenTelemetry, observability, SRE
Languages
Go, TypeScript, Python, Java, C#, C
Leadership
Technical strategy, architecture review, cross-org initiatives, incident command, mentoring senior engineers

Contact

What’s not working?

A few lines about the problem are enough to get started.

Book a 30-minute call[email protected]